I started Synop in September 1998 as a uni student working from my bedroom with about $8,000 to my name. Luckily, starting a business didn’t seem like a big deal because I could always get a job and I didn’t really even think of it as starting a business, let alone understand what that meant.
Initially, Synop traded as a NSW business called Synop Software.
Caught up in the Internet gold rush, I wanted to build an online version of Quicken so I could enter my expenses remotely. Typical of Synop, this turned into a full fledged electronic organiser with sophisticated item sharing between members. elec.org was the working name and the site was built using AOLServer, TCL and the Solid database server. Having no experience with Internet sites or databases, Philip and Alex’s Guide to Web Publishing was my bible at this time.
In March 1999, Dave Thomas became involved as a mentor and guide on my business building journey. His probing questions led me to understand more about the business realities of running a large scale web site funded by advertising, particularly since disk space was relatively expensive. Around this time sites like When.com launched the first online electronic organisers.
During 1999, Dave found me various small paid consulting projects for his network. Only later did I realise that this was a gentle form of testing, funding and keeping me afloat while I learnt the basics of a business (i.e. one needs customers).
Around this time I shifted from AOLServer to using a PHP in a LAMP stack (although it wasn’t called that yet). The primary driver was that I could host a public web site with scripting and a database for $25US / month. Demonstrating elec.org was horrendously difficult and expensive as AOLserver was not commonly used and required a dedicated server.
Inspired by software that ArsDigita was building, I built a PHP equivalent called Synphony. This powered e-gineer.com, to which I was adding many articles and instruction sets and which hosted the PHP Knowledge Base (later FAQTs).
In August 1999 I made my first visit to the US. Reading extensively online, I kept hearing about the internet revolution and how the web was everywhere. In Australia, the occaisional URL was making its way onto the bottom of TV advertisements and I thought “Yeah, it’s happening”. Then I arrived in San Francisco. Every billboard I saw during my visit was for an Internet company. I could literally overhear people talking about new business ideas across cafes. I learnt two things: the significant divide between Australia and the US, despite the Internet; and just how huge the Internet was becoming.
I did a number of job interviews in the US on this trip, and decided to turn down jobs with ArsDigita, CNet and others. It’s at this point that Synop really became more than a hobby.
Towards the end of that trip (October), I ended up in Sarasota, Florida upgrading WorldFinanceNet.com from FrontPage and static HTML to run the first version of Synphony. This top 1000 web site with massive morning traffic spikes was Synop’s first real customer and user of Synphony.
WFN was a classic Internet boom company. Two decisions stand out in particular. The first was deciding not to spend $5,000 US to get the wfn.com domain name for their top 1000 web site (apparently, WorldFinanceNet.com really isn’t that long or confusing). The second was when senior management came in and told me with much excitement that we need to add another advertisement to the home page. It was a barter deal in exchange for cheap access to their own private jet for getting to meetings…
At this point, the realities of web server support became very real to me with many late night calls. Being woken up to cries of “the server’s down!” is not conducive to a good nights sleep. One night, I logged in and saw 300+ httpd processes thrashing on the server. Restarting Apache, I couldn’t believe my eyes as the traffic built up to 200+ again within a minute. Turns out that a hacker at WFN had changed numerous includes on the home page to pull in smaller pieces on the site using URLs rather than PHP file includes. The result, was a morning traffic spike with every home page visit spawning 5 PHP script requests to Apache.
Unfortunately, the WFN relationship didn’t end happily (Apr 2000). Business lessons: bill customers early and often; it’s hard to be persistent and insistent by phone; and the US legal system is too expensive to bother pursuing $50k US ($100k AU) in unpaid bills.
There was one alternative I didn’t pursue however. On a US domestic flight I ended up next to a crazy old duck from NY. Turns out she was a bookmaker in her spare time and was more than happy to give me the number of some guys who could be quite persuasive when retrieving funds.
During the summer of 99/00, Jad and I built FAQTs.com and really took Synphony to the next level. The highlight was probably teaching Jad how to kick a football in my living room over many intense design discussions. At this point Synop had it’s own room in our apartment for an office.
Apachecon 2000 was held in Orlando during March, at the absolute peak of the bubble. Everyone was running around offering everyone else a job. This was the first of a number of conferences where I was lucky enough to present.
July 2000 was a dark time for Synop, with its survival limited to my credit cards after the WFN payment default. Through Randy Best, a great friend to Synop, we entered a contract with First Light Communications in New York to build and maintain a content management system that they would resell. “Feast or famine” was an email subject from Randy at the time, and certainly securing this contract took Synop back into the fast lane.
Synop Software incorporated to become Synop Pty Ltd at this time. Luckily another company with the name Synop had dropped it between 1998 and 2000, so we could get the originally preferred name.
Over the coming months I built and delivered Synphony v2 products to First Light. The real crunch was just prior to the Sydney olympics when I did three 20-hour days in a row, shipping the first product on the day of the opening ceremony. I remember working to 3am and getting up at 6:30am for a chance to watch the torch run by on the Pacific Hwy. Synphony obtained role based security, membership models, affiliate tracking and an incredible raft of features at this time.
In October 2000, my good friend Matt joined as Synop’s first employee. He quickly established our Artarmon office, which we used until closing and generally built all of Synop’s administration facilities and systems that I’d completely ignored since starting. John and Dean joined Synop quickly afterwards, ramping up our scale to support Synphony and continue development.
Hopes and expectations for the future of Synop were high, unrealistically high on my part, around this time. In anticipation of a bright future, the Synphony products were rebranded to Sytadel and a suite of E-gineers (communication, community, construction, content, knowledge, surveillance, sytadel).
These hopes came crashing back to earth when we held a booth at ApacheCon 2001. This was a different world to just a year before, with few attendees and everyone there looking for a job rather than offering them. The bubble had truly burst.
Over the coming months all of Synop’s significant clients went out of business, and Synop was forced to shrink back down to me working alone. At this stage I was experimenting with tools to help with Agile project management (code name Agilation), including an add-in for Microsoft project that provided a card playing UI for task allocation.
Peter Bailey had joined as a part time consultant in May 2001 and agreed to join permanently after helping to win our first government contract (the IP Access portal) in Oct 2001. Thus began the next phase in Synop’s life, and a great journey for us together.
Sytadel v3 was built and shipped as part of the IP Access project. This saw version control, workflow and many other advanced features come to Sytadel. At this point, Sytadel owners could use the Construction E-gineer to develop new content types that were automatically version controlled, subject to workflow, had complete editing and viewing screens. All the generated code was completely documented and extensively commented for ease of editing.
I completed the final stages of this project working from Internet cafes while chasing Bianca around Central America. Actually, the fastest Internet connection I’ve ever used was in Guatemala.
In April 2002, Synop organised Agile Australia 2002, the first series of talks on Agile Methods in Oz. I also presented two Synop papers at XP 2002 in Italy. To reflect this change and our project development methods, Synop’s tagline became “take the agile approach”.
Mid-2002 saw Synop win the ACCC Internet and Intranet portal redevelopment project in partnership with SecureNet. This was a huge project with more than 750 pages in the project specification. Many project problems, including massive scope creep, performance bottlenecks and my decision to rewrite Sytadel from scratch into v4 based on XML saw this project blow out into 2004. As a fixed price concern, this project became a huge millstone around Synop’s neck.
Although I did have my car written off while parked outside the ACCC late one night, the darkest days of this project were spent working for a week at the James Court Serviced Apartments in Canberra. After months of furious code development and sleepless nights, we felt Sytadel had reached a turning point in solidity and features. This week was going to put on the final touches and really begin roll out of the myriad of ACCC customisations. But, PHP and Sablotron had other plans for us. At this point, the system literally imploded and everything just stopped working. Days of installs, work arounds, bug fixes and rollbacks could not get things working again. Sytadel is an incredibly complex piece of engineering that pushes PHP, MySQL and XML libraries to their limit. Or beyond it, as we found out that week. After returning to Sydney (broken shadows of our former selves), we did some C hacking to replace the XSLT engine and Sytadel magically restored to working order. I still get shivers every time I pass the James Court, but only Richard who joined Synop and toiled with me through every inch of the Sytadel v4 rewrite would probably understand that.
Buried in the ACCC debacle, Synop struggled through early 2003 to deliver on a range of new customer projects (e.g. UNJLC, MDBC) we’d won with the promise of Sytadel v4 features and hard business development work throughout 2002. Working on site at the ACCC and in Peter’s Canberra home office, Matt Sheppard joined us in late 2002 to help.
Through 2003, Synop succesfully added hosting, support and consulting revenue streams to smooth out our traditionally lumpy revenue.
Contemplating life after this project onslaught, in April 2003 we started thinking about an AusIndustry R&D START grant application for what eventually became the Sauce project. Work on Sauce began in September 2003, at which point Victor Hadianto also joined Synop.
The Sauce project had 3 components: an aggregation server (TrustedSauce.com), aggregation client (Sauce Reader) and a content server (Sauce Studio). We built the aggregation server using .NET in late 2003 but it was never made public due to lack of resources to support a large search infrastructure and compete with now well-funded players like Technorati. TrustedSauce contained unique features for FOAF discovery, tracing weblog conversations and more.
Sauce Reader was released as a free for personal use product, to support RSS reading and weblog posting. Many 1.x versions were released through 2004, using the .NET platform. To address significant performance concerns, and for better integration with Sauce Studio, Sauce Reader 2.x was built using Delphi and released in May 2005. While technically advanced, Sauce Reader struggled as the RSS reader market exploded through 2004 with hundreds of competitors, free open source versions and increasing integration into web browsers.
Sauce Studio was built to concept stage in house and provided a file system based, XML content management system. It was a lightweight, flexible solution that could be combined with third party security, version control and other tools. Although we had strong belief in this technology, unfortunately we could not identify an appropriate market niche to commercialise the product.
While this aggressive R&D took place we continued growing the Sytadel and consulting aspects to Synop’s business. We opened our Canberra office in January, and by late 2004, Synop had grown into a company that could estimate and deliver increasingly complex projects reliably and profitably (e.g. METeOR, AER). During this busy time, Synop doubled in size to 10 people, adding more developers and our first full time business development manager. We finally had a good mix of experience and skill-sets.
Come 2005, Synop was faced with a difficult situation. Despite growing our business development resources and stringing together successful projects, our pipeline of new Sytadel business was running out. Our projects were typically the more high risk and complex customisation jobs requiring a flexible CMS. But, hypercompetition was driving prices lower and making jobs increasingly difficult for Synop to win.
After recognising the inherent conflict of being a small company building large enterprise CMS systems back in 2003, the Sauce project was intended to provide the next phase of Synop’s product line and consulting services. However, delays due to hangover projects in 2004 and a failure to identify an immediate and profitable market niche made rapid growth of revenue from this project highly unlikely.
With Synop in a financially healthy, but strategically weak situation we considered a range of alternatives for the future of the business. The only financially viable option appeared to be movement towards a pure consulting services model, building on third party products. Unfortunately, Synop’s passion had always been developing new products. So, the sad and difficult decision was made to divest assets and close the business as appropriate.